To address the issues of frequent identity switches(IDs)and degraded identification accuracy in multi object tracking(MOT)under complex occlusion scenarios,this study proposes an occlusion-robust tracking framework ba...To address the issues of frequent identity switches(IDs)and degraded identification accuracy in multi object tracking(MOT)under complex occlusion scenarios,this study proposes an occlusion-robust tracking framework based on face-pedestrian joint feature modeling.By constructing a joint tracking model centered on“intra-class independent tracking+cross-category dynamic binding”,designing a multi-modal matching metric with spatio-temporal and appearance constraints,and innovatively introducing a cross-category feature mutual verification mechanism and a dual matching strategy,this work effectively resolves performance degradation in traditional single-category tracking methods caused by short-term occlusion,cross-camera tracking,and crowded environments.Experiments on the Chokepoint_Face_Pedestrian_Track test set demonstrate that in complex scenes,the proposed method improves Face-Pedestrian Matching F1 area under the curve(F1 AUC)by approximately 4 to 43 percentage points compared to several traditional methods.The joint tracking model achieves overall performance metrics of IDF1:85.1825%and MOTA:86.5956%,representing improvements of 0.91 and 0.06 percentage points,respectively,over the baseline model.Ablation studies confirm the effectiveness of key modules such as the Intersection over Area(IoA)/Intersection over Union(IoU)joint metric and dynamic threshold adjustment,validating the significant role of the cross-category identity matching mechanism in enhancing tracking stability.Our_model shows a 16.7%frame per second(FPS)drop vs.fairness of detection and re-identification in multiple object tracking(FairMOT),with its cross-category binding module adding aboute 10%overhead,yet maintains near-real-time performance for essential face-pedestrian tracking at small resolutions.展开更多
Automated essay scoring(AES)systems have gained significant importance in educational settings,offering a scalable,efficient,and objective method for evaluating student essays.However,developing AES systems for Arabic...Automated essay scoring(AES)systems have gained significant importance in educational settings,offering a scalable,efficient,and objective method for evaluating student essays.However,developing AES systems for Arabic poses distinct challenges due to the language’s complex morphology,diglossia,and the scarcity of annotated datasets.This paper presents a hybrid approach to Arabic AES by combining text-based,vector-based,and embeddingbased similarity measures to improve essay scoring accuracy while minimizing the training data required.Using a large Arabic essay dataset categorized into thematic groups,the study conducted four experiments to evaluate the impact of feature selection,data size,and model performance.Experiment 1 established a baseline using a non-machine learning approach,selecting top-N correlated features to predict essay scores.The subsequent experiments employed 5-fold cross-validation.Experiment 2 showed that combining embedding-based,text-based,and vector-based features in a Random Forest(RF)model achieved an R2 of 88.92%and an accuracy of 83.3%within a 0.5-point tolerance.Experiment 3 further refined the feature selection process,demonstrating that 19 correlated features yielded optimal results,improving R2 to 88.95%.In Experiment 4,an optimal data efficiency training approach was introduced,where training data portions increased from 5%to 50%.The study found that using just 10%of the data achieved near-peak performance,with an R2 of 85.49%,emphasizing an effective trade-off between performance and computational costs.These findings highlight the potential of the hybrid approach for developing scalable Arabic AES systems,especially in low-resource environments,addressing linguistic challenges while ensuring efficient data usage.展开更多
Feature selection(FS)is a pivotal pre-processing step in developing data-driven models,influencing reliability,performance and optimization.Although existing FS techniques can yield high-performance metrics for certai...Feature selection(FS)is a pivotal pre-processing step in developing data-driven models,influencing reliability,performance and optimization.Although existing FS techniques can yield high-performance metrics for certain models,they do not invariably guarantee the extraction of the most critical or impactful features.Prior literature underscores the significance of equitable FS practices and has proposed diverse methodologies for the identification of appropriate features.However,the challenge of discerning the most relevant and influential features persists,particularly in the context of the exponential growth and heterogeneity of big data—a challenge that is increasingly salient in modern artificial intelligence(AI)applications.In response,this study introduces an innovative,automated statistical method termed Farea Similarity for Feature Selection(FSFS).The FSFS approach computes a similarity metric for each feature by benchmarking it against the record-wise mean,thereby finding feature dependencies and mitigating the influence of outliers that could potentially distort evaluation outcomes.Features are subsequently ranked according to their similarity scores,with the threshold established at the average similarity score.Notably,lower FSFS values indicate higher similarity and stronger data correlations,whereas higher values suggest lower similarity.The FSFS method is designed not only to yield reliable evaluation metrics but also to reduce data complexity without compromising model performance.Comparative analyses were performed against several established techniques,including Chi-squared(CS),Correlation Coefficient(CC),Genetic Algorithm(GA),Exhaustive Approach,Greedy Stepwise Approach,Gain Ratio,and Filtered Subset Eval,using a variety of datasets such as the Experimental Dataset,Breast Cancer Wisconsin(Original),KDD CUP 1999,NSL-KDD,UNSW-NB15,and Edge-IIoT.In the absence of the FSFS method,the highest classifier accuracies observed were 60.00%,95.13%,97.02%,98.17%,95.86%,and 94.62%for the respective datasets.When the FSFS technique was integrated with data normalization,encoding,balancing,and feature importance selection processes,accuracies improved to 100.00%,97.81%,98.63%,98.94%,94.27%,and 98.46%,respectively.The FSFS method,with a computational complexity of O(fn log n),demonstrates robust scalability and is well-suited for datasets of large size,ensuring efficient processing even when the number of features is substantial.By automatically eliminating outliers and redundant data,FSFS reduces computational overhead,resulting in faster training and improved model performance.Overall,the FSFS framework not only optimizes performance but also enhances the interpretability and explainability of data-driven models,thereby facilitating more trustworthy decision-making in AI applications.展开更多
Detecting cyber attacks in networks connected to the Internet of Things(IoT)is of utmost importance because of the growing vulnerabilities in the smart environment.Conventional models,such as Naive Bayes and support v...Detecting cyber attacks in networks connected to the Internet of Things(IoT)is of utmost importance because of the growing vulnerabilities in the smart environment.Conventional models,such as Naive Bayes and support vector machine(SVM),as well as ensemble methods,such as Gradient Boosting and eXtreme gradient boosting(XGBoost),are often plagued by high computational costs,which makes it challenging for them to perform real-time detection.In this regard,we suggested an attack detection approach that integrates Visual Geometry Group 16(VGG16),Artificial Rabbits Optimizer(ARO),and Random Forest Model to increase detection accuracy and operational efficiency in Internet of Things(IoT)networks.In the suggested model,the extraction of features from malware pictures was accomplished with the help of VGG16.The prediction process is carried out by the random forest model using the extracted features from the VGG16.Additionally,ARO is used to improve the hyper-parameters of the random forest model of the random forest.With an accuracy of 96.36%,the suggested model outperforms the standard models in terms of accuracy,F1-score,precision,and recall.The comparative research highlights our strategy’s success,which improves performance while maintaining a lower computational cost.This method is ideal for real-time applications,but it is effective.展开更多
Selecting proper descriptors(also known feature selection,FS)is key in the process of establishing mechanical properties prediction model of hot-rolled microalloyed steels by using machine learning(ML)algorithm.FS met...Selecting proper descriptors(also known feature selection,FS)is key in the process of establishing mechanical properties prediction model of hot-rolled microalloyed steels by using machine learning(ML)algorithm.FS methods based on data-driving can reduce the redundancy of data features and improve the prediction accuracy of mechanical properties.Based on the collected data of hot-rolled microalloyed steels,the association rules are used to mine the correlation information between the data.High-quality feature subsets are selected by the proposed FS method(FS method based on genetic algorithm embedding,GAMIC).Compared with the common FS method,it is shown on dataset that GAMIC selects feature subsets more appropriately.Six different ML algorithms are trained and tested for mechanical properties prediction.The result shows that the root-mean-square error of yield strength,tensile strength and elongation based on limit gradient enhancement(XGBoost)algorithm is 21.95 MPa,20.85 MPa and 1.96%,the correlation coefficient(R^(2))is 0.969,0.968 and 0.830,and the mean absolute error is 16.84 MPa,15.83 MPa and 1.48%,respectively,showing the best prediction performance.Finally,SHapley Additive exPlanation is used to further explore the influence of feature variables on mechanical properties.GAMIC feature selection method proposed is universal,which provides a basis for the development of high-precision mechanical property prediction model.展开更多
The authors regret that the original publication of this paper did not include Jawad Fayaz as a co-author.After further discussions and a thorough review of the research contributions,it was agreed that his significan...The authors regret that the original publication of this paper did not include Jawad Fayaz as a co-author.After further discussions and a thorough review of the research contributions,it was agreed that his significant contributions to the foundational aspects of the research warranted recognition,and he has now been added as a co-author.展开更多
In endoscopic surgery,the limited field of view and the nonlinear deformation of organs caused by patient movement and respiration significantly complicate the modeling and accurate tracking of soft tissue surfaces fr...In endoscopic surgery,the limited field of view and the nonlinear deformation of organs caused by patient movement and respiration significantly complicate the modeling and accurate tracking of soft tissue surfaces from endoscopic image sequences.To address these challenges,we propose a novel Hybrid Triangular Matching(HTM)modeling framework for soft tissue feature tracking.Specifically,HTM constructs a geometric model of the detected blobs on the soft tissue surface by applying the Watershed algorithm for blob detection and integrating the Delaunay triangulation with a newly designed triangle search segmentation algorithm.By leveraging barycentric coordinate theory,HTMrapidly and accurately establishes inter-frame correspondences within the triangulated model,enabling stable feature tracking without explicit markers or extensive training data.Experimental results on endoscopic sequences demonstrate that this model-based tracking approach achieves lower computational complexity,maintains robustness against tissue deformation,and provides a scalable geometric modeling method for real-time soft tissue tracking in surgical computer vision.展开更多
In this paper,a feature selection method for determining input parameters in antenna modeling is proposed.In antenna modeling,the input feature of artificial neural network(ANN)is geometric parameters.The selection cr...In this paper,a feature selection method for determining input parameters in antenna modeling is proposed.In antenna modeling,the input feature of artificial neural network(ANN)is geometric parameters.The selection criteria contain correlation and sensitivity between the geometric parameter and the electromagnetic(EM)response.Maximal information coefficient(MIC),an exploratory data mining tool,is introduced to evaluate both linear and nonlinear correlations.The EM response range is utilized to evaluate the sensitivity.The wide response range corresponding to varying values of a parameter implies the parameter is highly sensitive and the narrow response range suggests the parameter is insensitive.Only the parameter which is highly correlative and sensitive is selected as the input of ANN,and the sampling space of the model is highly reduced.The modeling of a wideband and circularly polarized antenna is studied as an example to verify the effectiveness of the proposed method.The number of input parameters decreases from8 to 4.The testing errors of|S_(11)|and axis ratio are reduced by8.74%and 8.95%,respectively,compared with the ANN with no feature selection.展开更多
Accurate prediction of landslide displacement is crucial for effective early warning of landslide disasters.While most existing prediction methods focus on time-series forecasting for individual monitoring points,ther...Accurate prediction of landslide displacement is crucial for effective early warning of landslide disasters.While most existing prediction methods focus on time-series forecasting for individual monitoring points,there is limited research on the spatiotemporal characteristics of landslide deformation.This paper proposes a novel Multi-Relation Spatiotemporal Graph Residual Network with Multi-Level Feature Attention(MFA-MRSTGRN)that effectively improves the prediction performance of landslide displacement through spatiotemporal fusion.This model integrates internal seepage factors as data feature enhancements with external triggering factors,allowing for accurate capture of the complex spatiotemporal characteristics of landslide displacement and the construction of a multi-source heterogeneous dataset.The MFA-MRSTGRN model incorporates dynamic graph theory and four key modules:multilevel feature attention,temporal-residual decomposition,spatial multi-relational graph convolution,and spatiotemporal fusion prediction.This comprehensive approach enables the efficient analyses of multi-source heterogeneous datasets,facilitating adaptive exploration of the evolving multi-relational,multi-dimensional spatiotemporal complexities in landslides.When applying this model to predict the displacement of the Liangshuijing landslide,we demonstrate that the MFA-MRSTGRN model surpasses traditional models,such as random forest(RF),long short-term memory(LSTM),and spatial temporal graph convolutional networks(ST-GCN)models in terms of various evaluation metrics including mean absolute error(MAE=1.27 mm),root mean square error(RMSE=1.49 mm),mean absolute percentage error(MAPE=0.026),and R-squared(R^(2)=0.88).Furthermore,feature ablation experiments indicate that incorporating internal seepage factors improves the predictive performance of landslide displacement models.This research provides an advanced and reliable method for landslide displacement prediction.展开更多
Aiming at the problem of low detection accuracy due to the different scale sizes of apple leaf disease spots and their similarity to the background,this paper proposes a multi-scale lightweight network(MSL-Net).Firstl...Aiming at the problem of low detection accuracy due to the different scale sizes of apple leaf disease spots and their similarity to the background,this paper proposes a multi-scale lightweight network(MSL-Net).Firstly,a multiplexed aggregated feature extraction network is proposed using residual bottleneck block(RES-Bottleneck)and middle partial-convolution(MP-Conv)to capture multi-scale spatial features and enhance focus on disease features for better differentiation between disease targets and background information.Secondly,a lightweight feature fusion network is designed using scale-fuse concatenation(SF-Cat)and triple-scale sequence feature fusion(TSSF)module to merge multi-scale feature maps comprehensively.Depthwise convolution(DWConv)and GhostNet lighten the network,while the cross stage partial bottleneck with 3 convolutions ghost-normalization attention module(C3-GN)reduces missed detections by suppressing irrelevant background information.Finally,soft non-maximum suppression(Soft-NMS)is used in the post-processing stage to improve the problem of misdetection of dense disease sites.The results show that the MSL-Net improves mean average precision at intersection over union of 0.5(mAP@0.5)by 2.0%over the baseline you only look once version 5s(YOLOv5s)and reduces parameters by 44%,reducing computation by 27%,outperforming other state-of-the-art(SOTA)models overall.This method also shows excellent performance compared to the latest research.展开更多
Predictive maintenance plays a crucial role in preventing equipment failures and minimizing operational downtime in modern industries.However,traditional predictive maintenance methods often face challenges in adaptin...Predictive maintenance plays a crucial role in preventing equipment failures and minimizing operational downtime in modern industries.However,traditional predictive maintenance methods often face challenges in adapting to diverse industrial environments and ensuring the transparency and fairness of their predictions.This paper presents a novel predictive maintenance framework that integrates deep learning and optimization techniques while addressing key ethical considerations,such as transparency,fairness,and explainability,in artificial intelligence driven decision-making.The framework employs an Autoencoder for feature reduction,a Convolutional Neural Network for pattern recognition,and a Long Short-Term Memory network for temporal analysis.To enhance transparency,the decision-making process of the framework is made interpretable,allowing stakeholders to understand and trust the model’s predictions.Additionally,Particle Swarm Optimization is used to refine hyperparameters for optimal performance and mitigate potential biases in the model.Experiments are conducted on multiple datasets from different industrial scenarios,with performance validated using accuracy,precision,recall,F1-score,and training time metrics.The results demonstrate an impressive accuracy of up to 99.92%and 99.45%across different datasets,highlighting the framework’s effectiveness in enhancing predictive maintenance strategies.Furthermore,the model’s explainability ensures that the decisions can be audited for fairness and accountability,aligning with ethical standards for critical systems.By addressing transparency and reducing potential biases,this framework contributes to the responsible and trustworthy deployment of artificial intelligence in industrial environments,particularly in safety-critical applications.The results underscore its potential for wide application across various industrial contexts,enhancing both performance and ethical decision-making.展开更多
Relation extraction plays a crucial role in numerous downstream tasks.Dialogue relation extraction focuses on identifying relations between two arguments within a given dialogue.To tackle the problem of low informatio...Relation extraction plays a crucial role in numerous downstream tasks.Dialogue relation extraction focuses on identifying relations between two arguments within a given dialogue.To tackle the problem of low information density in dialogues,methods based on trigger enhancement have been proposed,yielding positive results.However,trigger enhancement faces challenges,which cause suboptimal model performance.First,the proportion of annotated triggers is low in DialogRE.Second,feature representations of triggers and arguments often contain conflicting information.In this paper,we propose a novel Multi-Feature Filtering and Fusion trigger enhancement approach to overcome these limitations.We first obtain representations of arguments,and triggers that contain rich semantic information through attention and gate methods.Then,we design a feature filtering mechanism that eliminates conflicting features in the encoding of trigger prototype representations and their corresponding argument pairs.Additionally,we utilize large language models to create prompts based on Chain-of-Thought and In-context Learning for automated trigger extraction.Experiments show that our model increases the average F1 score by 1.3%in the dialogue relation extraction task.Ablation and case studies confirm the effectiveness of our model.Furthermore,the feature filtering method effectively integrates with other trigger enhancement models,enhancing overall performance and demonstrating its ability to resolve feature conflicts.展开更多
Heart disease prediction is a critical issue in healthcare,where accurate early diagnosis can save lives and reduce healthcare costs.The problem is inherently complex due to the high dimensionality of medical data,irr...Heart disease prediction is a critical issue in healthcare,where accurate early diagnosis can save lives and reduce healthcare costs.The problem is inherently complex due to the high dimensionality of medical data,irrelevant or redundant features,and the variability in risk factors such as age,lifestyle,andmedical history.These challenges often lead to inefficient and less accuratemodels.Traditional predictionmethodologies face limitations in effectively handling large feature sets and optimizing classification performance,which can result in overfitting poor generalization,and high computational cost.This work proposes a novel classification model for heart disease prediction that addresses these challenges by integrating feature selection through a Genetic Algorithm(GA)with an ensemble deep learning approach optimized using the Tunicate Swarm Algorithm(TSA).GA selects the most relevant features,reducing dimensionality and improvingmodel efficiency.Theselected features are then used to train an ensemble of deep learning models,where the TSA optimizes the weight of each model in the ensemble to enhance prediction accuracy.This hybrid approach addresses key challenges in the field,such as high dimensionality,redundant features,and classification performance,by introducing an efficient feature selection mechanism and optimizing the weighting of deep learning models in the ensemble.These enhancements result in a model that achieves superior accuracy,generalization,and efficiency compared to traditional methods.The proposed model demonstrated notable advancements in both prediction accuracy and computational efficiency over traditionalmodels.Specifically,it achieved an accuracy of 97.5%,a sensitivity of 97.2%,and a specificity of 97.8%.Additionally,with a 60-40 data split and 5-fold cross-validation,the model showed a significant reduction in training time(90 s),memory consumption(950 MB),and CPU usage(80%),highlighting its effectiveness in processing large,complex medical datasets for heart disease prediction.展开更多
Objective To explore the feasibility of constructing a lung cancer early-warning risk model based on facial image features,providing novel insights into the early screening of lung cancer.Methods This study included p...Objective To explore the feasibility of constructing a lung cancer early-warning risk model based on facial image features,providing novel insights into the early screening of lung cancer.Methods This study included patients with pulmonary nodules diagnosed at the Physical Examination Center of Shuguang Hospital Affiliated to Shanghai University of Traditional Chinese Medicine from November 1,2019 to December 31,2024,as well as patients with lung cancer diagnosed in the Oncology Departments of Yueyang Hospital of Integrated Traditional Chinese and Western Medicine and Longhua Hospital during the same period.The facial image information of patients with pulmonary nodules and lung cancer was collected using the TFDA-1 tongue and facial diagnosis instrument,and the facial diagnosis features were extracted from it by deep learning technology.Statistical analysis was conducted on the objective facial diagnosis characteristics of the two groups of participants to explore the differences in their facial image characteristics,and the least absolute shrinkage and selection operator(LASSO)regression was used to screen the characteristic variables.Based on the screened feature variables,four machine learning methods:random forest,logistic regression,support vector machine(SVM),and gradient boosting decision tree(GBDT)were used to establish lung cancer classification models independently.Meanwhile,the model performance was evaluated by indicators such as sensitivity,specificity,F1 score,precision,accuracy,the area under the receiver operating characteristic(ROC)curve(AUC),and the area under the precision-recall curve(AP).Results A total of 1275 patients with pulmonary nodules and 1623 patients with lung cancer were included in this study.After propensity score matching(PSM)to adjust for gender and age,535 patients were finally included in the pulmonary nodule group and the lung cancer group,respectively.There were significant differences in multiple color space metrics(such as R,G,B,V,L,a,b,Cr,H,Y,and Cb)and texture metrics[such as gray-levcl co-occurrence matrix(GLCM)-contrast(CON)and GLCM-inverse different moment(IDM)]between the two groups of individuals with pulmonary nodules and lung cancer(P<0.05).To construct a classification model,LASSO regression was used to select 63 key features from the initial 136 facial features.Based on this feature set,the SVM model demonstrated the best performance after 10-fold stratified cross-validation.The model achieved an average AUC of 0.8729 and average accuracy of 0.7990 on the internal test set.Further validation on an independent test set confirmed the model’s robust performance(AUC=0.8233,accuracy=0.7290),indicating its good generalization ability.Feature importance analysis demonstrated that color space indicators and the whole/lip Cr components(including color-B-0,wholecolor-Cr,and lipcolor-Cr)were the core factors in the model’s classification decisions,while texture indicators[GLCM-angular second moment(ASM)_2,GLCM-IDM_1,GLCM-CON_1,GLCM-entropy(ENT)_2]played an important auxiliary role.Conclusion The facial image features of patients with lung cancer and pulmonary nodules show significant differences in color and texture characteristics in multiple areas.The various models constructed based on facial image features all demonstrate good performance,indicating that facial image features can serve as potential biomarkers for lung cancer risk prediction,providing a non-invasive and feasible new approach for early lung cancer screening.展开更多
The detection of Ochotona Curzoniae serves as a fundamental component for estimating the population size of this species and for analyzing the dynamics of its population fluctuations.In natural environments,the pixels...The detection of Ochotona Curzoniae serves as a fundamental component for estimating the population size of this species and for analyzing the dynamics of its population fluctuations.In natural environments,the pixels representing Ochotona Curzoniae constitute a small fraction of the total pixels,and their distinguishing features are often subtle,complicating the target detection process.To effectively extract the characteristics of these small targets,a feature fusion approach that utilizes up-sampling and channel integration from various layers within a CNN can significantly enhance the representation of target features,ultimately improving detection accuracy.However,the top-down fusion of features from different layers may lead to information duplication and semantic bias,resulting in redundancy and high-frequency noise.To address the challenges of information redundancy and high-frequency noise during the feature fusion process in CNN,we have developed a target detection model for Ochotona Curzoniae.This model is based on a spatial-channel reconfiguration convolutional(SCConv)attentional mechanism and feature fusion(FFBCA),integrated with the Faster R-CNN framework.It consists of a feature extraction network,an attention mechanism-based feature fusion module,and a jump residual connection fusion module.Initially,we designed a dual attention mechanism feature fusion module that employs spatial-channel reconstruction convolution.In the spatial dimension,the attention mechanism adopts a separation-reconstruction approach,calculating a weight matrix for the spatial information within the feature map through group normalization.This process directs the model to concentrate on feature information assigned varying weights,thereby reducing redundancy during feature fusion.In the channel dimension,the attention mechanism utilizes a partition-transpose-fusion method,segmenting the input feature map into high-noise and low-noise components based on the variance of the feature information.The high-noise segment is processed through a low-pass filter constructed from pointwise convolution(PWC)to eliminate some high-frequency noise,while the low-noise segment employs a bottleneck structure with global average pooling(GAP)to generate a weight matrix that emphasizes the significance of channel dimension feature information.This approach diminishes the model’s focus on low-weight feature information,thereby preserving low-frequency semantic information while reducing information redundancy.Furthermore,we have developed a novel feature extraction network,ResNeXt-S,by integrating the Sim attention mechanism into ResNeXt50.This configuration assigns three-dimensional attention weights to each position within the feature map,thereby enhancing the local feature information of small targets while reducing background noise.Finally,we constructed a jump residual connection fusion module to minimize the loss of high-level semantic information during the feature fusion process.Experiments on Ochotona Curzoniae target detection on the Ochotona Curzoniae dataset show that the detection accuracy of the model in this paper is 92.3%,which is higher than that of FSSD512(84.6%),TDFSSD512(81.3%),FPN(86.5%),FFBAM(88.5%),Faster R-CNN(89.6%),and SSD512(88.6%)detection accuracies.展开更多
Efficient road distress detection is crucial for transportation safety.To address the challenge of balancing detection accuracy,efficiency,and multi-scale feature fusion in existing methods,this paper proposes a light...Efficient road distress detection is crucial for transportation safety.To address the challenge of balancing detection accuracy,efficiency,and multi-scale feature fusion in existing methods,this paper proposes a lightweight model named MADF-YOLOv8.The model enhances multi-scale feature extraction capability by introducing the Multi-Scale Ghost Residual Convolution(MSGRConv)and the Multiscale Adaptive Feature Processing Module(MAFP).Furthermore,it constructs a Multi-scale Dynamic sampling Bidirectional Feature Pyramid Network(MD-BiFPN)and incorporates the C2f-Faster module to optimize feature fusion efficiency.Experiments on the RDD2022 dataset demonstrate that the proposed model achieves a mean Average Precision at 0.5 Intersection over Union(mAP@0.5)of 88.6%with only 2.312 million parameters.Its overall performance surpasses various mainstream detectors,achieving an exceptional balance between accuracy and efficiency.展开更多
Different synthetic aperture radar(SAR)sensors vary significantly in resolution,polarization modes,and frequency bands,making it difficult to directly apply existing models to newly launched SAR satellites.These new s...Different synthetic aperture radar(SAR)sensors vary significantly in resolution,polarization modes,and frequency bands,making it difficult to directly apply existing models to newly launched SAR satellites.These new systems require large amounts of labeled data for model retraining,but collecting sufficient data in a short time is often infeasible.To address this contradiction,this paper proposes a data generation and transfer framework,integrating a stable diffusion model with attention distillation,that leverages historical SAR data to synthesize training data tailored to the unique characteristics of new SAR systems.Specifically,we fine-tune the low-rank adaptation(LoRA)modules within the multimodal diffusion transformer(MM-DiT)architecture to enable class-controllable SAR image generation guided by textual prompts.To ensure that the generated images reflect the statistical properties and imaging characteristics of the target SAR system,we further introduce an attention distillation mechanism that transfers sensor-specific features,such as spatial texture,speckle distribution,and structural patterns,from real target-domain data to the generative model.Extensive experiments on multi-class aircraft target datasets from two real spaceborne SAR systems demonstrate the effectiveness of the proposed approach in alleviating data scarcity and supporting cross-sensor remote sensing applications.展开更多
Urban parks are essential habitats for birds in densely built environments.However,most studies focus only on horizontal features,overlooking the role of integrated vertical structures.Using bird records from the Chin...Urban parks are essential habitats for birds in densely built environments.However,most studies focus only on horizontal features,overlooking the role of integrated vertical structures.Using bird records from the China Bird Report(CBR) collected between 2020 and 2024,this study focused on major urban parks in Harbin,a cold highlatitude city in Northeast China.Using interpretable machine learning models,we assessed the impact of threedimensional(3D) park features on bird species richness and identified key variables along with their nonlinear relationships.Subsequently,we grouped bird species by ecological traits and applied ordination analysis to assess their differential abundance responses to the 3D environmental characteristics of urban parks.Results showed that:(1) urban park 3D environmental features had a certain influence on bird richness,among which disturbance from built-up land and the coverage of commercial buildings were the most important factors;(2) different bird groups responded differently to the urban park 3D environmental features.Specifically,waterbirds and generalist birds preferred mixed habitats dominated by water,wetlands,and tall vegetation,and exhibited significant edge effects caused by built-up land disturbance.In contrast,forest birds were mainly affected by greenspace configuration.Although they showed tolerance to human disturbance,this trend may increase the risk of biotic homogenization.The study used a 3D environmental framework integrating horizontal and vertical dimensions to explore how bird diversity is shaped,providing detailed recommendations for biodiversitysensitive park planning in Harbin.展开更多
Multiple Sclerosis(MS)poses significant health risks.Patients may face neurodegeneration,mobility issues,cognitive decline,and a reduced quality of life.Manual diagnosis by neurologists is prone to limitations,making ...Multiple Sclerosis(MS)poses significant health risks.Patients may face neurodegeneration,mobility issues,cognitive decline,and a reduced quality of life.Manual diagnosis by neurologists is prone to limitations,making AI-based classification crucial for early detection.Therefore,automated classification using Artificial Intelligence(AI)techniques has a crucial role in addressing the limitations of manual classification and preventing the development of MS to advanced stages.This study developed hybrid systems integrating XGBoost(eXtreme Gradient Boosting)with multi-CNN(Convolutional Neural Networks)features based on Ant Colony Optimization(ACO)and Maximum Entropy Score-based Selection(MESbS)algorithms for early classification of MRI(Magnetic Resonance Imaging)images in a multi-class and binary-class MS dataset.All hybrid systems started by enhancing MRI images using the fusion processes of a Gaussian filter and Contrast-Limited Adaptive Histogram Equalization(CLAHE).Then,the Gradient Vector Flow(GVF)algorithm was applied to select white matter(regions of interest)within the brain and segment them from the surrounding brain structures.These regions of interest were processed by CNN models(ResNet101,DenseNet201,and MobileNet)to extract deep feature maps,which were then combined into fused feature vectors of multi-CNN model combinations(ResNet101-DenseNet201,DenseNet201-MobileNet,ResNet101-MobileNet,and ResNet101-DenseNet201-MobileNet).The multi-CNN features underwent dimensionality reduction using ACO and MESbS algorithms to remove unimportant features and retain important features.The XGBoost classifier employed the resultant feature vectors for classification.All developed hybrid systems displayed promising outcomes.For multiclass classification,the XGBoost model using ResNet101-DenseNet201-MobileNet features selected by ACO attained 99.4%accuracy,99.45%precision,and 99.75%specificity,surpassing prior studies(93.76%accuracy).It reached 99.6%accuracy,99.65%precision,and 99.55%specificity in binary-class classification.These results demonstrate the effectiveness of multi-CNN fusion with feature selection in improving MS classification accuracy.展开更多
Brain tumor classification is crucial for personalized treatment planning.Although deep learning-based Artificial Intelligence(AI)models can automatically analyze tumor images,fine details of small tumor regions may b...Brain tumor classification is crucial for personalized treatment planning.Although deep learning-based Artificial Intelligence(AI)models can automatically analyze tumor images,fine details of small tumor regions may be overlooked during global feature extraction.Therefore,we propose a brain tumor Magnetic Resonance Imaging(MRI)classification model based on a global-local parallel dual-branch structure.The global branch employs ResNet50 with a Multi-Head Self-Attention(MHSA)to capture global contextual information from whole brain images,while the local branch utilizes VGG16 to extract fine-grained features from segmented brain tumor regions.The features from both branches are processed through designed attention-enhanced feature fusion module to filter and integrate important features.Additionally,to address sample imbalance in the dataset,we introduce a category attention block to improve the recognition of minority classes.Experimental results indicate that our method achieved a classification accuracy of 98.04%and a micro-average Area Under the Curve(AUC)of 0.989 in the classification of three types of brain tumors,surpassing several existing pre-trained Convolutional Neural Network(CNN)models.Additionally,feature interpretability analysis validated the effectiveness of the proposed model.This suggests that the method holds significant potential for brain tumor image classification.展开更多
基金supported by the confidential research grant No.a8317。
文摘To address the issues of frequent identity switches(IDs)and degraded identification accuracy in multi object tracking(MOT)under complex occlusion scenarios,this study proposes an occlusion-robust tracking framework based on face-pedestrian joint feature modeling.By constructing a joint tracking model centered on“intra-class independent tracking+cross-category dynamic binding”,designing a multi-modal matching metric with spatio-temporal and appearance constraints,and innovatively introducing a cross-category feature mutual verification mechanism and a dual matching strategy,this work effectively resolves performance degradation in traditional single-category tracking methods caused by short-term occlusion,cross-camera tracking,and crowded environments.Experiments on the Chokepoint_Face_Pedestrian_Track test set demonstrate that in complex scenes,the proposed method improves Face-Pedestrian Matching F1 area under the curve(F1 AUC)by approximately 4 to 43 percentage points compared to several traditional methods.The joint tracking model achieves overall performance metrics of IDF1:85.1825%and MOTA:86.5956%,representing improvements of 0.91 and 0.06 percentage points,respectively,over the baseline model.Ablation studies confirm the effectiveness of key modules such as the Intersection over Area(IoA)/Intersection over Union(IoU)joint metric and dynamic threshold adjustment,validating the significant role of the cross-category identity matching mechanism in enhancing tracking stability.Our_model shows a 16.7%frame per second(FPS)drop vs.fairness of detection and re-identification in multiple object tracking(FairMOT),with its cross-category binding module adding aboute 10%overhead,yet maintains near-real-time performance for essential face-pedestrian tracking at small resolutions.
基金funded by Deanship of Graduate studies and Scientific Research at Jouf University under grant No.(DGSSR-2024-02-01264).
文摘Automated essay scoring(AES)systems have gained significant importance in educational settings,offering a scalable,efficient,and objective method for evaluating student essays.However,developing AES systems for Arabic poses distinct challenges due to the language’s complex morphology,diglossia,and the scarcity of annotated datasets.This paper presents a hybrid approach to Arabic AES by combining text-based,vector-based,and embeddingbased similarity measures to improve essay scoring accuracy while minimizing the training data required.Using a large Arabic essay dataset categorized into thematic groups,the study conducted four experiments to evaluate the impact of feature selection,data size,and model performance.Experiment 1 established a baseline using a non-machine learning approach,selecting top-N correlated features to predict essay scores.The subsequent experiments employed 5-fold cross-validation.Experiment 2 showed that combining embedding-based,text-based,and vector-based features in a Random Forest(RF)model achieved an R2 of 88.92%and an accuracy of 83.3%within a 0.5-point tolerance.Experiment 3 further refined the feature selection process,demonstrating that 19 correlated features yielded optimal results,improving R2 to 88.95%.In Experiment 4,an optimal data efficiency training approach was introduced,where training data portions increased from 5%to 50%.The study found that using just 10%of the data achieved near-peak performance,with an R2 of 85.49%,emphasizing an effective trade-off between performance and computational costs.These findings highlight the potential of the hybrid approach for developing scalable Arabic AES systems,especially in low-resource environments,addressing linguistic challenges while ensuring efficient data usage.
文摘Feature selection(FS)is a pivotal pre-processing step in developing data-driven models,influencing reliability,performance and optimization.Although existing FS techniques can yield high-performance metrics for certain models,they do not invariably guarantee the extraction of the most critical or impactful features.Prior literature underscores the significance of equitable FS practices and has proposed diverse methodologies for the identification of appropriate features.However,the challenge of discerning the most relevant and influential features persists,particularly in the context of the exponential growth and heterogeneity of big data—a challenge that is increasingly salient in modern artificial intelligence(AI)applications.In response,this study introduces an innovative,automated statistical method termed Farea Similarity for Feature Selection(FSFS).The FSFS approach computes a similarity metric for each feature by benchmarking it against the record-wise mean,thereby finding feature dependencies and mitigating the influence of outliers that could potentially distort evaluation outcomes.Features are subsequently ranked according to their similarity scores,with the threshold established at the average similarity score.Notably,lower FSFS values indicate higher similarity and stronger data correlations,whereas higher values suggest lower similarity.The FSFS method is designed not only to yield reliable evaluation metrics but also to reduce data complexity without compromising model performance.Comparative analyses were performed against several established techniques,including Chi-squared(CS),Correlation Coefficient(CC),Genetic Algorithm(GA),Exhaustive Approach,Greedy Stepwise Approach,Gain Ratio,and Filtered Subset Eval,using a variety of datasets such as the Experimental Dataset,Breast Cancer Wisconsin(Original),KDD CUP 1999,NSL-KDD,UNSW-NB15,and Edge-IIoT.In the absence of the FSFS method,the highest classifier accuracies observed were 60.00%,95.13%,97.02%,98.17%,95.86%,and 94.62%for the respective datasets.When the FSFS technique was integrated with data normalization,encoding,balancing,and feature importance selection processes,accuracies improved to 100.00%,97.81%,98.63%,98.94%,94.27%,and 98.46%,respectively.The FSFS method,with a computational complexity of O(fn log n),demonstrates robust scalability and is well-suited for datasets of large size,ensuring efficient processing even when the number of features is substantial.By automatically eliminating outliers and redundant data,FSFS reduces computational overhead,resulting in faster training and improved model performance.Overall,the FSFS framework not only optimizes performance but also enhances the interpretability and explainability of data-driven models,thereby facilitating more trustworthy decision-making in AI applications.
基金funded by Institutional Fund Projects under grant no.(IFPDP-261-22)。
文摘Detecting cyber attacks in networks connected to the Internet of Things(IoT)is of utmost importance because of the growing vulnerabilities in the smart environment.Conventional models,such as Naive Bayes and support vector machine(SVM),as well as ensemble methods,such as Gradient Boosting and eXtreme gradient boosting(XGBoost),are often plagued by high computational costs,which makes it challenging for them to perform real-time detection.In this regard,we suggested an attack detection approach that integrates Visual Geometry Group 16(VGG16),Artificial Rabbits Optimizer(ARO),and Random Forest Model to increase detection accuracy and operational efficiency in Internet of Things(IoT)networks.In the suggested model,the extraction of features from malware pictures was accomplished with the help of VGG16.The prediction process is carried out by the random forest model using the extracted features from the VGG16.Additionally,ARO is used to improve the hyper-parameters of the random forest model of the random forest.With an accuracy of 96.36%,the suggested model outperforms the standard models in terms of accuracy,F1-score,precision,and recall.The comparative research highlights our strategy’s success,which improves performance while maintaining a lower computational cost.This method is ideal for real-time applications,but it is effective.
基金supported by the National Key Research and Development Program of China(Grant No.2021YFB3702404)the National Natural Science Foundation of China(Grant No.52104370)+4 种基金the Reviving-Liaoning Excellence Plan(XLYC2203186)Science and Technology Special Projects of Liaoning Province(Grant No.2022JH25/10200001)the Postdoctoral Research Fund for Northeastern(Grant No.20210203)Independent Projects of Basic Scientific Research(ZZ2021005)CITIC Niobium Steel Development Award Fund(2022-M1824).
文摘Selecting proper descriptors(also known feature selection,FS)is key in the process of establishing mechanical properties prediction model of hot-rolled microalloyed steels by using machine learning(ML)algorithm.FS methods based on data-driving can reduce the redundancy of data features and improve the prediction accuracy of mechanical properties.Based on the collected data of hot-rolled microalloyed steels,the association rules are used to mine the correlation information between the data.High-quality feature subsets are selected by the proposed FS method(FS method based on genetic algorithm embedding,GAMIC).Compared with the common FS method,it is shown on dataset that GAMIC selects feature subsets more appropriately.Six different ML algorithms are trained and tested for mechanical properties prediction.The result shows that the root-mean-square error of yield strength,tensile strength and elongation based on limit gradient enhancement(XGBoost)algorithm is 21.95 MPa,20.85 MPa and 1.96%,the correlation coefficient(R^(2))is 0.969,0.968 and 0.830,and the mean absolute error is 16.84 MPa,15.83 MPa and 1.48%,respectively,showing the best prediction performance.Finally,SHapley Additive exPlanation is used to further explore the influence of feature variables on mechanical properties.GAMIC feature selection method proposed is universal,which provides a basis for the development of high-precision mechanical property prediction model.
文摘The authors regret that the original publication of this paper did not include Jawad Fayaz as a co-author.After further discussions and a thorough review of the research contributions,it was agreed that his significant contributions to the foundational aspects of the research warranted recognition,and he has now been added as a co-author.
基金Support by Sichuan Science and Technology Program[2023YFSY0026,2023YFH0004].
文摘In endoscopic surgery,the limited field of view and the nonlinear deformation of organs caused by patient movement and respiration significantly complicate the modeling and accurate tracking of soft tissue surfaces from endoscopic image sequences.To address these challenges,we propose a novel Hybrid Triangular Matching(HTM)modeling framework for soft tissue feature tracking.Specifically,HTM constructs a geometric model of the detected blobs on the soft tissue surface by applying the Watershed algorithm for blob detection and integrating the Delaunay triangulation with a newly designed triangle search segmentation algorithm.By leveraging barycentric coordinate theory,HTMrapidly and accurately establishes inter-frame correspondences within the triangulated model,enabling stable feature tracking without explicit markers or extensive training data.Experimental results on endoscopic sequences demonstrate that this model-based tracking approach achieves lower computational complexity,maintains robustness against tissue deformation,and provides a scalable geometric modeling method for real-time soft tissue tracking in surgical computer vision.
基金National Natural Science Foundation of China(62161048)Sichuan Science and Technology Program(2022NSFSC0547,2022ZYD0109)。
文摘In this paper,a feature selection method for determining input parameters in antenna modeling is proposed.In antenna modeling,the input feature of artificial neural network(ANN)is geometric parameters.The selection criteria contain correlation and sensitivity between the geometric parameter and the electromagnetic(EM)response.Maximal information coefficient(MIC),an exploratory data mining tool,is introduced to evaluate both linear and nonlinear correlations.The EM response range is utilized to evaluate the sensitivity.The wide response range corresponding to varying values of a parameter implies the parameter is highly sensitive and the narrow response range suggests the parameter is insensitive.Only the parameter which is highly correlative and sensitive is selected as the input of ANN,and the sampling space of the model is highly reduced.The modeling of a wideband and circularly polarized antenna is studied as an example to verify the effectiveness of the proposed method.The number of input parameters decreases from8 to 4.The testing errors of|S_(11)|and axis ratio are reduced by8.74%and 8.95%,respectively,compared with the ANN with no feature selection.
基金the funding support from the National Natural Science Foundation of China(Grant No.52308340)Chongqing Talent Innovation and Entrepreneurship Demonstration Team Project(Grant No.cstc2024ycjh-bgzxm0012)the Science and Technology Projects supported by China Coal Technology and Engineering Chongqing Design and Research Institute(Group)Co.,Ltd.(Grant No.H20230317).
文摘Accurate prediction of landslide displacement is crucial for effective early warning of landslide disasters.While most existing prediction methods focus on time-series forecasting for individual monitoring points,there is limited research on the spatiotemporal characteristics of landslide deformation.This paper proposes a novel Multi-Relation Spatiotemporal Graph Residual Network with Multi-Level Feature Attention(MFA-MRSTGRN)that effectively improves the prediction performance of landslide displacement through spatiotemporal fusion.This model integrates internal seepage factors as data feature enhancements with external triggering factors,allowing for accurate capture of the complex spatiotemporal characteristics of landslide displacement and the construction of a multi-source heterogeneous dataset.The MFA-MRSTGRN model incorporates dynamic graph theory and four key modules:multilevel feature attention,temporal-residual decomposition,spatial multi-relational graph convolution,and spatiotemporal fusion prediction.This comprehensive approach enables the efficient analyses of multi-source heterogeneous datasets,facilitating adaptive exploration of the evolving multi-relational,multi-dimensional spatiotemporal complexities in landslides.When applying this model to predict the displacement of the Liangshuijing landslide,we demonstrate that the MFA-MRSTGRN model surpasses traditional models,such as random forest(RF),long short-term memory(LSTM),and spatial temporal graph convolutional networks(ST-GCN)models in terms of various evaluation metrics including mean absolute error(MAE=1.27 mm),root mean square error(RMSE=1.49 mm),mean absolute percentage error(MAPE=0.026),and R-squared(R^(2)=0.88).Furthermore,feature ablation experiments indicate that incorporating internal seepage factors improves the predictive performance of landslide displacement models.This research provides an advanced and reliable method for landslide displacement prediction.
文摘Aiming at the problem of low detection accuracy due to the different scale sizes of apple leaf disease spots and their similarity to the background,this paper proposes a multi-scale lightweight network(MSL-Net).Firstly,a multiplexed aggregated feature extraction network is proposed using residual bottleneck block(RES-Bottleneck)and middle partial-convolution(MP-Conv)to capture multi-scale spatial features and enhance focus on disease features for better differentiation between disease targets and background information.Secondly,a lightweight feature fusion network is designed using scale-fuse concatenation(SF-Cat)and triple-scale sequence feature fusion(TSSF)module to merge multi-scale feature maps comprehensively.Depthwise convolution(DWConv)and GhostNet lighten the network,while the cross stage partial bottleneck with 3 convolutions ghost-normalization attention module(C3-GN)reduces missed detections by suppressing irrelevant background information.Finally,soft non-maximum suppression(Soft-NMS)is used in the post-processing stage to improve the problem of misdetection of dense disease sites.The results show that the MSL-Net improves mean average precision at intersection over union of 0.5(mAP@0.5)by 2.0%over the baseline you only look once version 5s(YOLOv5s)and reduces parameters by 44%,reducing computation by 27%,outperforming other state-of-the-art(SOTA)models overall.This method also shows excellent performance compared to the latest research.
文摘Predictive maintenance plays a crucial role in preventing equipment failures and minimizing operational downtime in modern industries.However,traditional predictive maintenance methods often face challenges in adapting to diverse industrial environments and ensuring the transparency and fairness of their predictions.This paper presents a novel predictive maintenance framework that integrates deep learning and optimization techniques while addressing key ethical considerations,such as transparency,fairness,and explainability,in artificial intelligence driven decision-making.The framework employs an Autoencoder for feature reduction,a Convolutional Neural Network for pattern recognition,and a Long Short-Term Memory network for temporal analysis.To enhance transparency,the decision-making process of the framework is made interpretable,allowing stakeholders to understand and trust the model’s predictions.Additionally,Particle Swarm Optimization is used to refine hyperparameters for optimal performance and mitigate potential biases in the model.Experiments are conducted on multiple datasets from different industrial scenarios,with performance validated using accuracy,precision,recall,F1-score,and training time metrics.The results demonstrate an impressive accuracy of up to 99.92%and 99.45%across different datasets,highlighting the framework’s effectiveness in enhancing predictive maintenance strategies.Furthermore,the model’s explainability ensures that the decisions can be audited for fairness and accountability,aligning with ethical standards for critical systems.By addressing transparency and reducing potential biases,this framework contributes to the responsible and trustworthy deployment of artificial intelligence in industrial environments,particularly in safety-critical applications.The results underscore its potential for wide application across various industrial contexts,enhancing both performance and ethical decision-making.
基金supported by the National Key Research and Development Program of China(No.2023YFF0905400)the National Natural Science Foundation of China(No.U2341229).
文摘Relation extraction plays a crucial role in numerous downstream tasks.Dialogue relation extraction focuses on identifying relations between two arguments within a given dialogue.To tackle the problem of low information density in dialogues,methods based on trigger enhancement have been proposed,yielding positive results.However,trigger enhancement faces challenges,which cause suboptimal model performance.First,the proportion of annotated triggers is low in DialogRE.Second,feature representations of triggers and arguments often contain conflicting information.In this paper,we propose a novel Multi-Feature Filtering and Fusion trigger enhancement approach to overcome these limitations.We first obtain representations of arguments,and triggers that contain rich semantic information through attention and gate methods.Then,we design a feature filtering mechanism that eliminates conflicting features in the encoding of trigger prototype representations and their corresponding argument pairs.Additionally,we utilize large language models to create prompts based on Chain-of-Thought and In-context Learning for automated trigger extraction.Experiments show that our model increases the average F1 score by 1.3%in the dialogue relation extraction task.Ablation and case studies confirm the effectiveness of our model.Furthermore,the feature filtering method effectively integrates with other trigger enhancement models,enhancing overall performance and demonstrating its ability to resolve feature conflicts.
文摘Heart disease prediction is a critical issue in healthcare,where accurate early diagnosis can save lives and reduce healthcare costs.The problem is inherently complex due to the high dimensionality of medical data,irrelevant or redundant features,and the variability in risk factors such as age,lifestyle,andmedical history.These challenges often lead to inefficient and less accuratemodels.Traditional predictionmethodologies face limitations in effectively handling large feature sets and optimizing classification performance,which can result in overfitting poor generalization,and high computational cost.This work proposes a novel classification model for heart disease prediction that addresses these challenges by integrating feature selection through a Genetic Algorithm(GA)with an ensemble deep learning approach optimized using the Tunicate Swarm Algorithm(TSA).GA selects the most relevant features,reducing dimensionality and improvingmodel efficiency.Theselected features are then used to train an ensemble of deep learning models,where the TSA optimizes the weight of each model in the ensemble to enhance prediction accuracy.This hybrid approach addresses key challenges in the field,such as high dimensionality,redundant features,and classification performance,by introducing an efficient feature selection mechanism and optimizing the weighting of deep learning models in the ensemble.These enhancements result in a model that achieves superior accuracy,generalization,and efficiency compared to traditional methods.The proposed model demonstrated notable advancements in both prediction accuracy and computational efficiency over traditionalmodels.Specifically,it achieved an accuracy of 97.5%,a sensitivity of 97.2%,and a specificity of 97.8%.Additionally,with a 60-40 data split and 5-fold cross-validation,the model showed a significant reduction in training time(90 s),memory consumption(950 MB),and CPU usage(80%),highlighting its effectiveness in processing large,complex medical datasets for heart disease prediction.
基金National Natural Science Foundation of China(82305090)Shanghai Municipal Health Commission(20234Y0168)National Key Research and Development Program of China (2017YFC1703301)。
文摘Objective To explore the feasibility of constructing a lung cancer early-warning risk model based on facial image features,providing novel insights into the early screening of lung cancer.Methods This study included patients with pulmonary nodules diagnosed at the Physical Examination Center of Shuguang Hospital Affiliated to Shanghai University of Traditional Chinese Medicine from November 1,2019 to December 31,2024,as well as patients with lung cancer diagnosed in the Oncology Departments of Yueyang Hospital of Integrated Traditional Chinese and Western Medicine and Longhua Hospital during the same period.The facial image information of patients with pulmonary nodules and lung cancer was collected using the TFDA-1 tongue and facial diagnosis instrument,and the facial diagnosis features were extracted from it by deep learning technology.Statistical analysis was conducted on the objective facial diagnosis characteristics of the two groups of participants to explore the differences in their facial image characteristics,and the least absolute shrinkage and selection operator(LASSO)regression was used to screen the characteristic variables.Based on the screened feature variables,four machine learning methods:random forest,logistic regression,support vector machine(SVM),and gradient boosting decision tree(GBDT)were used to establish lung cancer classification models independently.Meanwhile,the model performance was evaluated by indicators such as sensitivity,specificity,F1 score,precision,accuracy,the area under the receiver operating characteristic(ROC)curve(AUC),and the area under the precision-recall curve(AP).Results A total of 1275 patients with pulmonary nodules and 1623 patients with lung cancer were included in this study.After propensity score matching(PSM)to adjust for gender and age,535 patients were finally included in the pulmonary nodule group and the lung cancer group,respectively.There were significant differences in multiple color space metrics(such as R,G,B,V,L,a,b,Cr,H,Y,and Cb)and texture metrics[such as gray-levcl co-occurrence matrix(GLCM)-contrast(CON)and GLCM-inverse different moment(IDM)]between the two groups of individuals with pulmonary nodules and lung cancer(P<0.05).To construct a classification model,LASSO regression was used to select 63 key features from the initial 136 facial features.Based on this feature set,the SVM model demonstrated the best performance after 10-fold stratified cross-validation.The model achieved an average AUC of 0.8729 and average accuracy of 0.7990 on the internal test set.Further validation on an independent test set confirmed the model’s robust performance(AUC=0.8233,accuracy=0.7290),indicating its good generalization ability.Feature importance analysis demonstrated that color space indicators and the whole/lip Cr components(including color-B-0,wholecolor-Cr,and lipcolor-Cr)were the core factors in the model’s classification decisions,while texture indicators[GLCM-angular second moment(ASM)_2,GLCM-IDM_1,GLCM-CON_1,GLCM-entropy(ENT)_2]played an important auxiliary role.Conclusion The facial image features of patients with lung cancer and pulmonary nodules show significant differences in color and texture characteristics in multiple areas.The various models constructed based on facial image features all demonstrate good performance,indicating that facial image features can serve as potential biomarkers for lung cancer risk prediction,providing a non-invasive and feasible new approach for early lung cancer screening.
基金funded by the National Natural Science Foundation of China(Grant Nos.62161019,62061024).
文摘The detection of Ochotona Curzoniae serves as a fundamental component for estimating the population size of this species and for analyzing the dynamics of its population fluctuations.In natural environments,the pixels representing Ochotona Curzoniae constitute a small fraction of the total pixels,and their distinguishing features are often subtle,complicating the target detection process.To effectively extract the characteristics of these small targets,a feature fusion approach that utilizes up-sampling and channel integration from various layers within a CNN can significantly enhance the representation of target features,ultimately improving detection accuracy.However,the top-down fusion of features from different layers may lead to information duplication and semantic bias,resulting in redundancy and high-frequency noise.To address the challenges of information redundancy and high-frequency noise during the feature fusion process in CNN,we have developed a target detection model for Ochotona Curzoniae.This model is based on a spatial-channel reconfiguration convolutional(SCConv)attentional mechanism and feature fusion(FFBCA),integrated with the Faster R-CNN framework.It consists of a feature extraction network,an attention mechanism-based feature fusion module,and a jump residual connection fusion module.Initially,we designed a dual attention mechanism feature fusion module that employs spatial-channel reconstruction convolution.In the spatial dimension,the attention mechanism adopts a separation-reconstruction approach,calculating a weight matrix for the spatial information within the feature map through group normalization.This process directs the model to concentrate on feature information assigned varying weights,thereby reducing redundancy during feature fusion.In the channel dimension,the attention mechanism utilizes a partition-transpose-fusion method,segmenting the input feature map into high-noise and low-noise components based on the variance of the feature information.The high-noise segment is processed through a low-pass filter constructed from pointwise convolution(PWC)to eliminate some high-frequency noise,while the low-noise segment employs a bottleneck structure with global average pooling(GAP)to generate a weight matrix that emphasizes the significance of channel dimension feature information.This approach diminishes the model’s focus on low-weight feature information,thereby preserving low-frequency semantic information while reducing information redundancy.Furthermore,we have developed a novel feature extraction network,ResNeXt-S,by integrating the Sim attention mechanism into ResNeXt50.This configuration assigns three-dimensional attention weights to each position within the feature map,thereby enhancing the local feature information of small targets while reducing background noise.Finally,we constructed a jump residual connection fusion module to minimize the loss of high-level semantic information during the feature fusion process.Experiments on Ochotona Curzoniae target detection on the Ochotona Curzoniae dataset show that the detection accuracy of the model in this paper is 92.3%,which is higher than that of FSSD512(84.6%),TDFSSD512(81.3%),FPN(86.5%),FFBAM(88.5%),Faster R-CNN(89.6%),and SSD512(88.6%)detection accuracies.
基金supported by the National Key Research and Development Program of China under Grant No.2022YFA1602003,entitled"Intelligent Monitoring of Taishan Neutrino Detector".
文摘Efficient road distress detection is crucial for transportation safety.To address the challenge of balancing detection accuracy,efficiency,and multi-scale feature fusion in existing methods,this paper proposes a lightweight model named MADF-YOLOv8.The model enhances multi-scale feature extraction capability by introducing the Multi-Scale Ghost Residual Convolution(MSGRConv)and the Multiscale Adaptive Feature Processing Module(MAFP).Furthermore,it constructs a Multi-scale Dynamic sampling Bidirectional Feature Pyramid Network(MD-BiFPN)and incorporates the C2f-Faster module to optimize feature fusion efficiency.Experiments on the RDD2022 dataset demonstrate that the proposed model achieves a mean Average Precision at 0.5 Intersection over Union(mAP@0.5)of 88.6%with only 2.312 million parameters.Its overall performance surpasses various mainstream detectors,achieving an exceptional balance between accuracy and efficiency.
基金supported in part by the National Natural Science Foundations of China(Nos.62201027,62271034)。
文摘Different synthetic aperture radar(SAR)sensors vary significantly in resolution,polarization modes,and frequency bands,making it difficult to directly apply existing models to newly launched SAR satellites.These new systems require large amounts of labeled data for model retraining,but collecting sufficient data in a short time is often infeasible.To address this contradiction,this paper proposes a data generation and transfer framework,integrating a stable diffusion model with attention distillation,that leverages historical SAR data to synthesize training data tailored to the unique characteristics of new SAR systems.Specifically,we fine-tune the low-rank adaptation(LoRA)modules within the multimodal diffusion transformer(MM-DiT)architecture to enable class-controllable SAR image generation guided by textual prompts.To ensure that the generated images reflect the statistical properties and imaging characteristics of the target SAR system,we further introduce an attention distillation mechanism that transfers sensor-specific features,such as spatial texture,speckle distribution,and structural patterns,from real target-domain data to the generative model.Extensive experiments on multi-class aircraft target datasets from two real spaceborne SAR systems demonstrate the effectiveness of the proposed approach in alleviating data scarcity and supporting cross-sensor remote sensing applications.
基金supported by Natural Science Foundation of Heilongjiang Province of China [Grant number LH2023E007]。
文摘Urban parks are essential habitats for birds in densely built environments.However,most studies focus only on horizontal features,overlooking the role of integrated vertical structures.Using bird records from the China Bird Report(CBR) collected between 2020 and 2024,this study focused on major urban parks in Harbin,a cold highlatitude city in Northeast China.Using interpretable machine learning models,we assessed the impact of threedimensional(3D) park features on bird species richness and identified key variables along with their nonlinear relationships.Subsequently,we grouped bird species by ecological traits and applied ordination analysis to assess their differential abundance responses to the 3D environmental characteristics of urban parks.Results showed that:(1) urban park 3D environmental features had a certain influence on bird richness,among which disturbance from built-up land and the coverage of commercial buildings were the most important factors;(2) different bird groups responded differently to the urban park 3D environmental features.Specifically,waterbirds and generalist birds preferred mixed habitats dominated by water,wetlands,and tall vegetation,and exhibited significant edge effects caused by built-up land disturbance.In contrast,forest birds were mainly affected by greenspace configuration.Although they showed tolerance to human disturbance,this trend may increase the risk of biotic homogenization.The study used a 3D environmental framework integrating horizontal and vertical dimensions to explore how bird diversity is shaped,providing detailed recommendations for biodiversitysensitive park planning in Harbin.
文摘Multiple Sclerosis(MS)poses significant health risks.Patients may face neurodegeneration,mobility issues,cognitive decline,and a reduced quality of life.Manual diagnosis by neurologists is prone to limitations,making AI-based classification crucial for early detection.Therefore,automated classification using Artificial Intelligence(AI)techniques has a crucial role in addressing the limitations of manual classification and preventing the development of MS to advanced stages.This study developed hybrid systems integrating XGBoost(eXtreme Gradient Boosting)with multi-CNN(Convolutional Neural Networks)features based on Ant Colony Optimization(ACO)and Maximum Entropy Score-based Selection(MESbS)algorithms for early classification of MRI(Magnetic Resonance Imaging)images in a multi-class and binary-class MS dataset.All hybrid systems started by enhancing MRI images using the fusion processes of a Gaussian filter and Contrast-Limited Adaptive Histogram Equalization(CLAHE).Then,the Gradient Vector Flow(GVF)algorithm was applied to select white matter(regions of interest)within the brain and segment them from the surrounding brain structures.These regions of interest were processed by CNN models(ResNet101,DenseNet201,and MobileNet)to extract deep feature maps,which were then combined into fused feature vectors of multi-CNN model combinations(ResNet101-DenseNet201,DenseNet201-MobileNet,ResNet101-MobileNet,and ResNet101-DenseNet201-MobileNet).The multi-CNN features underwent dimensionality reduction using ACO and MESbS algorithms to remove unimportant features and retain important features.The XGBoost classifier employed the resultant feature vectors for classification.All developed hybrid systems displayed promising outcomes.For multiclass classification,the XGBoost model using ResNet101-DenseNet201-MobileNet features selected by ACO attained 99.4%accuracy,99.45%precision,and 99.75%specificity,surpassing prior studies(93.76%accuracy).It reached 99.6%accuracy,99.65%precision,and 99.55%specificity in binary-class classification.These results demonstrate the effectiveness of multi-CNN fusion with feature selection in improving MS classification accuracy.
文摘Brain tumor classification is crucial for personalized treatment planning.Although deep learning-based Artificial Intelligence(AI)models can automatically analyze tumor images,fine details of small tumor regions may be overlooked during global feature extraction.Therefore,we propose a brain tumor Magnetic Resonance Imaging(MRI)classification model based on a global-local parallel dual-branch structure.The global branch employs ResNet50 with a Multi-Head Self-Attention(MHSA)to capture global contextual information from whole brain images,while the local branch utilizes VGG16 to extract fine-grained features from segmented brain tumor regions.The features from both branches are processed through designed attention-enhanced feature fusion module to filter and integrate important features.Additionally,to address sample imbalance in the dataset,we introduce a category attention block to improve the recognition of minority classes.Experimental results indicate that our method achieved a classification accuracy of 98.04%and a micro-average Area Under the Curve(AUC)of 0.989 in the classification of three types of brain tumors,surpassing several existing pre-trained Convolutional Neural Network(CNN)models.Additionally,feature interpretability analysis validated the effectiveness of the proposed model.This suggests that the method holds significant potential for brain tumor image classification.
